Non-transitory computer-readable storage medium, data processing method, and data processing apparatus

Abstract

A non-transitory computer-readable storage medium storing a program that causes a processor to execute a process, the process includes reading a plurality of first solutions of an optimization problem represented by a combination of values of a plurality of state variables, converting the plurality of first solutions into a plurality of second solutions by executing principal component analysis on the plurality of first solutions, determining a region that indicates a spread of the plurality of second solutions in a solution space, generating a third solution located at a second position outside the region and away from a first position within the region by a first distance, converting the third solution into a fourth solution that consists of the plurality of state variables, and searching for a solution of the optimization problem by using the fourth solution as an initial value of the plurality of state variables.

What is claimed is: 
     
         1 . A non-transitory computer-readable storage medium storing a program that causes a processor included in a computer to execute a process, the process comprising:
 reading, from a memory that stores a plurality of first solutions of an optimization problem represented by a combination of values of a plurality of state variables included in an Ising-type evaluation function, the plurality of first solutions;   first converting the plurality of first solutions into a plurality of second solutions represented by a first number of variables smaller than a number of the plurality of state variables by executing principal component analysis on the plurality of first solutions;   determining a region that indicates a spread of the plurality of second solutions in a solution space that includes a dimension of the first number;   generating a third solution located at a second position outside the region and away from a first position within the region by a first distance;   second converting the third solution into a fourth solution that consists of the plurality of state variables; and   searching for a solution of the optimization problem by using the fourth solution as an initial value of the plurality of state variables.   
     
     
         2 . The non-transitory computer-readable storage medium according to  claim 1 , further comprising
 updating one of the plurality of first solutions by a fifth solution of the optimization problem obtained by the searching that uses the fourth solution as the initial value, and   repeating the reading, the first converting, the determining, the generating, the second converting, and the searching.   
     
     
         3 . The non-transitory computer-readable storage medium according to  claim 1 , wherein
 the region is determined based on positions of the plurality of second solutions in the solution space, and   the first position is a center point of the region.   
     
     
         4 . The non-transitory computer-readable storage medium according to  claim 1 ,
 wherein the second position away from the first position by the first distance is randomly determined in the solution space.   
     
     
         5 . A data processing method comprising:
 reading, from a memory that stores a plurality of first solutions of an optimization problem represented by a combination of values of a plurality of state variables included in an Ising-type evaluation function, the plurality of first solutions;   first converting the plurality of first solutions into a plurality of second solutions represented by a first number of variables smaller than a number of the plurality of state variables by executing principal component analysis on the plurality of first solutions;   determining a region that indicates a spread of the plurality of second solutions in a solution space that includes a dimension of the first number;   generating a third solution located at a second position outside the region and away from a first position within the region by a first distance;   second converting the third solution into a fourth solution that consists of the plurality of state variables; and   searching for a solution of the optimization problem by using the fourth solution as an initial value of the plurality of state variables.   
     
     
         6 . A data processing apparatus comprising:
 a memory configured to store a plurality of first solutions of an optimization problem represented by a combination of values of a plurality of state variables included in an Ising-type evaluation function; and   a processor configured to:   read, from the memory, the plurality of first solutions,   convert the plurality of first solutions into a plurality of second solutions represented by a first number of variables smaller than a number of the plurality of state variables by executing principal component analysis on the plurality of first solutions;   determine a region that indicates a spread of the plurality of second solutions in a solution space that includes a dimension of the first number;   generate a third solution located at a second position outside the region and away from a first position within the region by a first distance;   convert the third solution into a fourth solution that consists of the plurality of state variables; and   search for a solution of the optimization problem by using the fourth solution as an initial value of the plurality of state variables.